Symmetric multiprocessor

1. inst.eecs.berkeley.edu/~cs61cCS61C : Machine StructuresLecture 5 – C Memory Management2010-06-28 Instructor Paul Pearce Symmetric multiprocessor MIPS support for AndroidMIPS Technologies (founded by John Hennessy, author of your textbook!) recently announced support for SMP on Android. This week, you’ll learn MIPS! http://tinyurl.com/297ne2j

2. Review • Use handles to change pointers • Create abstractions (and your own data structures) with structures • Dynamically allocated heap memory must be manually deallocated in C. • Use malloc() and free() to allocate and de-allocate persistent storage.

3. How big are structs? • Recall C operator sizeof() which gives size in bytes (of type or variable) • How big is sizeof(p)? struct p { char x; int y;}; • 5 bytes? 8 bytes? • Compiler may word align integer y • Takeaway: Structs can be padded, use sizeof()

4. Don’t forget the globals! • Remember: • Structure declaration does not allocate memory • Variable declaration does allocate memory • So far we have talked about several different ways to allocate memory for data: • Declaration of a local variable int i; struct Node list; char *string; int ar[n]; • “Dynamic” allocation at runtime by calling allocation function (malloc). ptr = (struct Node *) malloc(sizeof(struct Node)*n); • One more possibility exists… • Data declared outside of any procedure (i.e., before main). • Similar to #1 above, but has “global” scope. int myGlobal; main() { }

5. C Memory Management • C has 3 pools of memory • Static storage: global variable storage, basically permanent, entire program run • The Stack: local variable storage, parameters, return address(location of “activation records” in Java or “stack frame” in C) • The Heap (dynamic malloc storage): data lives until deallocated by programmer • C requires knowing where objects are in memory, otherwise things don’t work as expected • Java hides location of objects

6. Normal C Memory Management ~ FFFF FFFFhex stack • A program’s address space contains 4 regions: • stack: local variables, grows downward • heap: space requested for pointers via malloc() ; resizes dynamically, grows upward • static data: variables declared outside main, does not grow or shrink • code: loaded when program starts, does not change heap static data code ~ 0hex For now, OS somehowprevents accesses between stack and heap (gray hash lines). Wait for virtual memory

7. Where are variables allocated? • If declared outside a procedure, allocated in “static” storage • If declared inside procedure, allocated on the “stack”and freed when procedure returns. • main() is a procedure int myGlobal; main() { int myTemp; }

8. frame frame frame The Stack • Stack frame includes: • Return “instruction” address • Parameters • Space for other local variables • Stack frames contiguous blocks of memory; stack pointer tells where bottom stack frame is • When procedure ends, stack frame is tossed off the stack; frees memory for future stack frames SP frame

9. Stack Pointer Stack Pointer Stack Pointer Stack Pointer Stack Pointer Stack • Last In, First Out (LIFO) data structure stack main () { a(0); } Stack grows down void a (int m) { b(1); } void b (int n) { c(2); } void c (int o) { d(3); } void d (int p) { }

10. SP SP SP fun()(y==3) printf()(y==?) main main main Who cares about stack management? • Pointers in C allow access to deallocated memory, leading to hard-to-find bugs ! int *fun() { int y; y = 3; return &y;};main () { int *stackAddr,content; stackAddr = fun(); content = *stackAddr;printf("%d", content);/* 3 */ content = *stackAddr;printf("%d", content);/*13451514 */};

11. The Heap (Dynamic memory) • Large pool of memory, not allocated in contiguous order • back-to-back requests for heap memory could result blocks very far apart • where Java new command allocates memory • In C, specify number of bytes of memory explicitly to allocate item int *ptr;ptr = (int *) malloc(sizeof(int));/* malloc returns type (void *),so need to cast to right type */ • malloc(): Allocates raw, uninitialized memory from heap

12. Memory Management • How do we manage memory? • Code, Static storage are easy: they never grow or shrink • Stack space is also easy: stack frames are created and destroyed in last-in, first-out (LIFO) order • Managing the heap is tricky:memory can be allocated / deallocated at any time

13. Heap Management Requirements • Want malloc() and free() to run quickly. • Want minimal memory overhead • Want to avoid fragmentation*– when most of our free memory is in many small chunks • In this case, we might have many free bytes but not be able to satisfy a large request since the free bytes are not contiguous in memory. * This is technically called external fragmention

14. R1 (100 bytes) R2 (1 byte) Heap Management • An example • Request R1 for 100 bytes • Request R2 for 1 byte • Memory from R1 is freed • Request R3 for 50 bytes • What if R3 was a request for 101 bytes?

15. R3? R3? Heap Management • An example • Request R1 for 100 bytes • Request R2 for 1 byte • Memory from R1 is freed • Request R3 for 50 bytes • What if R3 was a request for 101 bytes? R2 (1 byte)

16. Administrivia • You made it through the first week! Yay! • Don’t forget about the newsgroup! • Hasn’t been used as much as Paul would like. Paul has elaborated on some things from lecture on the newsgroup! • We start assembly tomorrow. Don’t forget about the reading • HW 2 due tomorrow at midnight. • Project 1 out now. Start early, lots of code to understand (and it’s really fun!) • Check back of handout for CS Illustrated on pointers and arrays!

17. More Administrivia! • Tentative schedule for the rest of the summer now online! • Plan accordingly! • Midterm currently slated for Thursday July 15th from 6pm-9pm • Does anyone have a hard conflict with this time? • Final time will be determined based on how the midterm goes at the scheduled time. The final will be on Thursday, August 12th • Pick up a copy of the cheating policy from the front. On it please write your name, login, and then sign it. You’ll need to give this to your Lab TA in order to be checked off for Lab 3!

18. K&R Malloc/Free Implementation • From Section 8.7 of K&R • Code in the book uses some C language features we haven’t discussed and is written in a very terse style, don’t worry if you can’t decipher the code • Each block of memory is preceded by a header that has two fields: size of the block and a pointer to the next block • All free blocks are kept in a circular linked list, the pointer field is unused in an allocated block

19. K&R Implementation • malloc() searches the free list for a block that is big enough. If none is found, more memory is requested from the operating system. If what it gets can’t satisfy the request, it fails. • free() checks if the blocks adjacent to the freed block are also free • If so, adjacent free blocks are merged (coalesced) into a single, larger free block • Otherwise, the freed block is just added to the free list

20. Choosing a block in malloc() • If there are multiple free blocks of memory that are big enough for some request, how do we choose which one to use? • best-fit: choose the smallest block that is big enough for the request • first-fit: choose the first block we see that is big enough • next-fit: like first-fit but remember where we finished searching and resume searching from there

21. Slab Allocator • A different approach to memory management (used in GNU libc) • Divide blocks in to “large” and “small” by picking an arbitrary threshold size. Blocks larger than this threshold are managed with a freelist (as before). • For small blocks, allocate blocks in sizes that are powers of 2 • e.g., if program wants to allocate 20 bytes, actually give it 32 bytes

22. Slab Allocator • Bookkeeping for small blocks is relatively easy: just use a bitmap for each range of blocks of the same size • Allocating is easy and fast: compute the size of the block to allocate and find a free bit in the corresponding bitmap. • Freeing is also easy and fast: figure out which slab the address belongs to and clear the corresponding bit.

23. Slab Allocator 16 byte blocks: 32 byte blocks: 64 byte blocks: 16 byte block bitmap: 11011000 32 byte block bitmap: 0111 64 byte block bitmap: 00

24. Slab Allocator Tradeoffs • Extremely fast for small blocks. • Slower for large blocks • But presumably the program will take more time to do something with a large block so the overhead is not as critical. • Minimal space overhead • No fragmentation (as we defined it before) for small blocks, but still have wasted space!

25. Internal vs. External Fragmentation • With the slab allocator, difference between requested size and next power of 2 is wasted • e.g., if program wants to allocate 20 bytes and we give it a 32 byte block, 12 bytes are unused. • We also refer to this as fragmentation, but call it internal fragmentation since the wasted space is actually within an allocated block. • External fragmentation: wasted space between allocated blocks.

26. Buddy System • Yet another memory management technique (used in Linux kernel) • Like GNU’s “slab allocator”, but only allocate blocks in sizes that are powers of 2 (internal fragmentation is possible) • Keep separate free lists for each size • e.g., separate free lists for 16 byte, 32 byte, 64 byte blocks, etc.

27. Buddy System • If no free block of size n is available, find a block of size 2n and split it in to two blocks of size n • When a block of size n is freed, if its neighbor of size n is also free, combine the blocks in to a single block of size 2n • Buddy is block in other half larger block • Same speed advantages as slab allocator buddies NOT buddies

28. Allocation Schemes • So which memory management scheme (K&R, slab, buddy) is best? • There is no single best approach for every application. • Different applications have different allocation / deallocation patterns. • A scheme that works well for one application may work poorly for another application.

29. Peer Instruction – Pros and Cons of fits • first-fit results in many small blocks at the beginning of the free list • next-fit is slower than first-fit, since it takes longer in steady state to find a match • best-fitleaves lots of tiny blocks 123 a) FFT b) FTT c) TFF d) TFT e) TTT

30. And in conclusion… • C has 3 pools of memory • Static storage: global variable storage, basically permanent, entire program run • The Stack: local variable storage, parameters, return address • The Heap (dynamic storage): malloc() grabs space from here, free() returns it. • malloc() handles free space with freelist. Three different ways to find free space when given a request: • First fit (find first one that’s free) • Next fit (same as first, but remembers where left off) • Best fit (finds most “snug” free space) • Internal vs External fragmentation!

31. Bonus slides • These are extra slides that used to be included in lecture notes, but have been moved to this, the “bonus” area to serve as a supplement. • The slides will appear in the order they would have in the normal presentation Bonus

32. Intel 80x86 C Memory Management • A C program’s 80x86 address space : • heap: space requested for pointers via malloc(); resizes dynamically, grows upward • static data: variables declared outside main, does not grow or shrink • code: loaded when program starts, does not change • stack: local variables, grows downward heap static data code ~ 08000000hex stack

33. Tradeoffs of allocation policies • Best-fit: Tries to limit fragmentation but at the cost of time (must examine all free blocks for each malloc). Leaves lots of small blocks (why?) • First-fit: Quicker than best-fit (why?) but potentially more fragmentation. Tends to concentrate small blocks at the beginning of the free list (why?) • Next-fit: Does not concentrate small blocks at front like first-fit, should be faster as a result.